Nonlinear Correction of LVDT Sensor Based on ACO-BP Neural Network
Minghui Li, Gong Qiangling, Wu Wenkai, Hai Hongyu, Ma Chenpei, Che Chang · Journal of Physics Conference Series · 2020
In this paper, the output signal of the linear variable differential transformer type displacement sensor (LVDT) is nonlinear, a BP neural network optimized by the ant colony algorithm is designed to fit and correct the nonlinear output of the LVDT. This scheme first uses the ant colony algorithm to search the optimal range of neural network weights and thresholds, and then uses the BP neural network to fit any non-linear function to LVDT nonlinear output fitting and correction, which overcomes the shortcomings of BP neural network easily falling into local minimum and slow convergence. Through MATLAB simulation experiments, it is concluded that the convergence speed, average error and average error percentage of the ACO-BP neural network are better than the BP neural network. This solution has certain feasibility for solving the LVDT nonlinear problem, and provides a new solution for the sensor nonlinear correction.